AI Impact - adding tooltip with the actual usage numbers

Release note

Introducing GitLab's AI Impact approach

A new approach to measuring AI effectiveness, with focus on business outcomes with Gitlab Value Stream Management.

Problem to solve

Currently the AI (Code Suggestions) Usage Rate, is display in %. Using percentages alone doesn't give users enough context to understand the data's scale or magnitude, so users might misinterpret it.

Related to these problems

Based on these Customer feedback

Proposal

Adding tooltip with the actual usage numbers - Unique Code Suggestions users / Total Monthly unique contributors:

Screenshot_2024-04-15_at_7.27.57

Implementation guide:

frontend - add tooltip

  • Update AI impact GraphQL query to retrieve codeContributorsCount and codeSuggestionsContributorsCount
  • Remove codeSuggestionsUsageRate from GraphQL query and instead calculate the rate using the new values (codeSuggestionsUsageRate = codeSuggestionsContributorsCount / codeContributorsCount)
  • Update AI impact table cells with a tooltip to display the new values

backend - follow up

  • Deprecate / remove the codeSuggestionsUsageRate from the GraphQL api
Edited by Haim Snir